Parent Directory | Revision Log
Revision 392 - (view) (download)
1 : | Isibaar | 333 | /************************************************************************** |
2 : | * | ||
3 : | * XVID MPEG-4 VIDEO CODEC | ||
4 : | * 8x8 block-based halfpel interpolation | ||
5 : | * | ||
6 : | * This program is free software; you can redistribute it and/or modify | ||
7 : | * it under the terms of the GNU General Public License as published by | ||
8 : | * the Free Software Foundation; either version 2 of the License, or | ||
9 : | * (at your option) any later version. | ||
10 : | * | ||
11 : | * This program is distributed in the hope that it will be useful, | ||
12 : | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
13 : | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
14 : | * GNU General Public License for more details. | ||
15 : | * | ||
16 : | * You should have received a copy of the GNU General Public License | ||
17 : | * along with this program; if not, write to the Free Software | ||
18 : | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. | ||
19 : | * | ||
20 : | *************************************************************************/ | ||
21 : | |||
22 : | /************************************************************************** | ||
23 : | * | ||
24 : | * History: | ||
25 : | * | ||
26 : | * 02.05.2002 add "interpolate8x8_c" for B-frame; MinChen <chenm002@163.com> | ||
27 : | * 27.12.2001 modified "compensate_halfpel" | ||
28 : | * 05.11.2001 initial version; (c)2001 peter ross <pross@cs.rmit.edu.au> | ||
29 : | * | ||
30 : | *************************************************************************/ | ||
31 : | |||
32 : | |||
33 : | #include "../portab.h" | ||
34 : | #include "interpolate8x8.h" | ||
35 : | |||
36 : | // function pointers | ||
37 : | INTERPOLATE8X8_PTR interpolate8x8_halfpel_h; | ||
38 : | INTERPOLATE8X8_PTR interpolate8x8_halfpel_v; | ||
39 : | INTERPOLATE8X8_PTR interpolate8x8_halfpel_hv; | ||
40 : | |||
41 : | |||
42 : | // dst = interpolate(src) | ||
43 : | |||
44 : | void | ||
45 : | interpolate8x8_halfpel_h_c(uint8_t * const dst, | ||
46 : | const uint8_t * const src, | ||
47 : | const uint32_t stride, | ||
48 : | const uint32_t rounding) | ||
49 : | { | ||
50 : | uint32_t i, j; | ||
51 : | |||
52 : | for (j = 0; j < 8; j++) { | ||
53 : | for (i = 0; i < 8; i++) { | ||
54 : | |||
55 : | int16_t tot = | ||
56 : | (int32_t) src[j * stride + i] + (int32_t) src[j * stride + i + | ||
57 : | 1]; | ||
58 : | |||
59 : | tot = (int32_t) ((tot + 1 - rounding) >> 1); | ||
60 : | dst[j * stride + i] = (uint8_t) tot; | ||
61 : | } | ||
62 : | } | ||
63 : | } | ||
64 : | |||
65 : | |||
66 : | |||
67 : | void | ||
68 : | interpolate8x8_halfpel_v_c(uint8_t * const dst, | ||
69 : | const uint8_t * const src, | ||
70 : | const uint32_t stride, | ||
71 : | const uint32_t rounding) | ||
72 : | { | ||
73 : | uint32_t i, j; | ||
74 : | |||
75 : | for (j = 0; j < 8; j++) { | ||
76 : | for (i = 0; i < 8; i++) { | ||
77 : | int16_t tot = src[j * stride + i] + src[j * stride + i + stride]; | ||
78 : | |||
79 : | tot = ((tot + 1 - rounding) >> 1); | ||
80 : | dst[j * stride + i] = (uint8_t) tot; | ||
81 : | } | ||
82 : | } | ||
83 : | } | ||
84 : | |||
85 : | |||
86 : | void | ||
87 : | interpolate8x8_halfpel_hv_c(uint8_t * const dst, | ||
88 : | const uint8_t * const src, | ||
89 : | const uint32_t stride, | ||
90 : | const uint32_t rounding) | ||
91 : | { | ||
92 : | uint32_t i, j; | ||
93 : | |||
94 : | for (j = 0; j < 8; j++) { | ||
95 : | for (i = 0; i < 8; i++) { | ||
96 : | int16_t tot = | ||
97 : | src[j * stride + i] + src[j * stride + i + 1] + | ||
98 : | src[j * stride + i + stride] + src[j * stride + i + stride + | ||
99 : | 1]; | ||
100 : | tot = ((tot + 2 - rounding) >> 2); | ||
101 : | dst[j * stride + i] = (uint8_t) tot; | ||
102 : | } | ||
103 : | } | ||
104 : | } | ||
105 : | |||
106 : | // add by MinChen <chenm001@163.com> | ||
107 : | // interpolate8x8 two pred block | ||
108 : | void | ||
109 : | interpolate8x8_c(uint8_t * const dst, | ||
110 : | const uint8_t * const src, | ||
111 : | const uint32_t x, | ||
112 : | const uint32_t y, | ||
113 : | const uint32_t stride) | ||
114 : | { | ||
115 : | uint32_t i, j; | ||
116 : | |||
117 : | for (j = 0; j < 8; j++) { | ||
118 : | for (i = 0; i < 8; i++) { | ||
119 : | int32_t tot = | ||
120 : | ((src[(y + j) * stride + x + i] + | ||
121 : | dst[(y + j) * stride + x + i] + 1) >> 1); | ||
122 : | dst[(y + j) * stride + x + i] = (uint8_t) tot; | ||
123 : | } | ||
124 : | } | ||
125 : | } | ||
126 : | |||
127 : | /************************************************************* | ||
128 : | * QPEL STUFF STARTS HERE * | ||
129 : | *************************************************************/ | ||
130 : | |||
131 : | #define CLIP(X,A,B) (X < A) ? (A) : ((X > B) ? (B) : (X)) | ||
132 : | |||
133 : | void interpolate8x8_lowpass_h(uint8_t *dst, uint8_t *src, int32_t dst_stride, int32_t src_stride, int32_t rounding) | ||
134 : | { | ||
135 : | int32_t i; | ||
136 : | |||
137 : | for(i = 0; i < 8; i++) | ||
138 : | { | ||
139 : | dst[0] = CLIP((((src[0] + src[1]) * 160 - (src[0] + src[2]) * 48 + (src[1] + src[3]) * 24 - (src[2] + src[4]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
140 : | dst[1] = CLIP((((src[1] + src[2]) * 160 - (src[0] + src[3]) * 48 + (src[0] + src[4]) * 24 - (src[1] + src[5]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
141 : | dst[2] = CLIP((((src[2] + src[3]) * 160 - (src[1] + src[4]) * 48 + (src[0] + src[5]) * 24 - (src[0] + src[6]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
142 : | dst[3] = CLIP((((src[3] + src[4]) * 160 - (src[2] + src[5]) * 48 + (src[1] + src[6]) * 24 - (src[0] + src[7]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
143 : | dst[4] = CLIP((((src[4] + src[5]) * 160 - (src[3] + src[6]) * 48 + (src[2] + src[7]) * 24 - (src[1] + src[8]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
144 : | dst[5] = CLIP((((src[5] + src[6]) * 160 - (src[4] + src[7]) * 48 + (src[3] + src[8]) * 24 - (src[2] + src[8]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
145 : | dst[6] = CLIP((((src[6] + src[7]) * 160 - (src[5] + src[8]) * 48 + (src[4] + src[8]) * 24 - (src[3] + src[7]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
146 : | dst[7] = CLIP((((src[7] + src[8]) * 160 - (src[6] + src[8]) * 48 + (src[5] + src[7]) * 24 - (src[4] + src[6]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
147 : | |||
148 : | dst += dst_stride; | ||
149 : | src += src_stride; | ||
150 : | } | ||
151 : | } | ||
152 : | |||
153 : | void interpolate8x8_lowpass_v(uint8_t *dst, uint8_t *src, int32_t dst_stride, int32_t src_stride, int32_t rounding) | ||
154 : | { | ||
155 : | int32_t i; | ||
156 : | |||
157 : | for(i = 0; i < 8; i++) | ||
158 : | { | ||
159 : | int32_t src0 = src[0]; | ||
160 : | int32_t src1 = src[src_stride]; | ||
161 : | int32_t src2 = src[2 * src_stride]; | ||
162 : | int32_t src3 = src[3 * src_stride]; | ||
163 : | int32_t src4 = src[4 * src_stride]; | ||
164 : | int32_t src5 = src[5 * src_stride]; | ||
165 : | int32_t src6 = src[6 * src_stride]; | ||
166 : | int32_t src7 = src[7 * src_stride]; | ||
167 : | int32_t src8 = src[8 * src_stride]; | ||
168 : | |||
169 : | dst[0] = CLIP((((src0 + src1) * 160 - (src0 + src2) * 48 + (src1 + src3) * 24 - (src2 + src4) * 8 + (128 - rounding)) / 256), 0, 255); | ||
170 : | dst[dst_stride] = CLIP((((src1 + src2) * 160 - (src0 + src3) * 48 + (src0 + src4) * 24 - (src1 + src5) * 8 + (128 - rounding)) / 256), 0, 255); | ||
171 : | dst[2 * dst_stride] = CLIP((((src2 + src3) * 160 - (src1 + src4) * 48 + (src0 + src5) * 24 - (src0 + src6) * 8 + (128 - rounding)) / 256), 0, 255); | ||
172 : | dst[3 * dst_stride] = CLIP((((src3 + src4) * 160 - (src2 + src5) * 48 + (src1 + src6) * 24 - (src0 + src7) * 8 + (128 - rounding)) / 256), 0, 255); | ||
173 : | dst[4 * dst_stride] = CLIP((((src4 + src5) * 160 - (src3 + src6) * 48 + (src2 + src7) * 24 - (src1 + src8) * 8 + (128 - rounding)) / 256), 0, 255); | ||
174 : | dst[5 * dst_stride] = CLIP((((src5 + src6) * 160 - (src4 + src7) * 48 + (src3 + src8) * 24 - (src2 + src8) * 8 + (128 - rounding)) / 256), 0, 255); | ||
175 : | dst[6 * dst_stride] = CLIP((((src6 + src7) * 160 - (src5 + src8) * 48 + (src4 + src8) * 24 - (src3 + src7) * 8 + (128 - rounding)) / 256), 0, 255); | ||
176 : | dst[7 * dst_stride] = CLIP((((src7 + src8) * 160 - (src6 + src8) * 48 + (src5 + src7) * 24 - (src4 + src6) * 8 + (128 - rounding)) / 256), 0, 255); | ||
177 : | |||
178 : | dst++; | ||
179 : | src++; | ||
180 : | } | ||
181 : | } | ||
182 : | |||
183 : | void interpolate8x8_lowpass_hv(uint8_t *dst1, uint8_t *dst2, uint8_t *src, int32_t dst1_stride, int32_t dst2_stride, int32_t src_stride, int32_t rounding) | ||
184 : | { | ||
185 : | uint8_t data[72]; | ||
186 : | |||
187 : | int32_t i; | ||
188 : | |||
189 : | |||
190 : | for(i = 0; i < 9; i++) | ||
191 : | { | ||
192 : | dst2[0] = data[8 * i + 0] = CLIP((((src[0] + src[1]) * 160 - (src[0] + src[2]) * 48 + (src[1] + src[3]) * 24 - (src[2] + src[4]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
193 : | dst2[1] = data[8 * i + 1] = CLIP((((src[1] + src[2]) * 160 - (src[0] + src[3]) * 48 + (src[0] + src[4]) * 24 - (src[1] + src[5]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
194 : | dst2[2] = data[8 * i + 2] = CLIP((((src[2] + src[3]) * 160 - (src[1] + src[4]) * 48 + (src[0] + src[5]) * 24 - (src[0] + src[6]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
195 : | dst2[3] = data[8 * i + 3] = CLIP((((src[3] + src[4]) * 160 - (src[2] + src[5]) * 48 + (src[1] + src[6]) * 24 - (src[0] + src[7]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
196 : | dst2[4] = data[8 * i + 4] = CLIP((((src[4] + src[5]) * 160 - (src[3] + src[6]) * 48 + (src[2] + src[7]) * 24 - (src[1] + src[8]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
197 : | dst2[5] = data[8 * i + 5] = CLIP((((src[5] + src[6]) * 160 - (src[4] + src[7]) * 48 + (src[3] + src[8]) * 24 - (src[2] + src[8]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
198 : | dst2[6] = data[8 * i + 6] = CLIP((((src[6] + src[7]) * 160 - (src[5] + src[8]) * 48 + (src[4] + src[8]) * 24 - (src[3] + src[7]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
199 : | dst2[7] = data[8 * i + 7] = CLIP((((src[7] + src[8]) * 160 - (src[6] + src[8]) * 48 + (src[5] + src[7]) * 24 - (src[4] + src[6]) * 8 + (128 - rounding)) / 256), 0, 255); | ||
200 : | |||
201 : | src += src_stride; | ||
202 : | dst2 += dst2_stride; | ||
203 : | } | ||
204 : | |||
205 : | for(i = 0; i < 8; i++) | ||
206 : | { | ||
207 : | int32_t src0 = data[i]; | ||
208 : | int32_t src1 = data[8 + i]; | ||
209 : | int32_t src2 = data[2 * 8 + i]; | ||
210 : | int32_t src3 = data[3 * 8 + i]; | ||
211 : | int32_t src4 = data[4 * 8 + i]; | ||
212 : | int32_t src5 = data[5 * 8 + i]; | ||
213 : | int32_t src6 = data[6 * 8 + i]; | ||
214 : | int32_t src7 = data[7 * 8 + i]; | ||
215 : | int32_t src8 = data[8 * 8 + i]; | ||
216 : | |||
217 : | dst1[0] = CLIP((((src0 + src1) * 160 - (src0 + src2) * 48 + (src1 + src3) * 24 - (src2 + src4) * 8 + (128 - rounding)) / 256), 0, 255); | ||
218 : | dst1[dst1_stride] = CLIP((((src1 + src2) * 160 - (src0 + src3) * 48 + (src0 + src4) * 24 - (src1 + src5) * 8 + (128 - rounding)) / 256), 0, 255); | ||
219 : | dst1[2 * dst1_stride] = CLIP((((src2 + src3) * 160 - (src1 + src4) * 48 + (src0 + src5) * 24 - (src0 + src6) * 8 + (128 - rounding)) / 256), 0, 255); | ||
220 : | dst1[3 * dst1_stride] = CLIP((((src3 + src4) * 160 - (src2 + src5) * 48 + (src1 + src6) * 24 - (src0 + src7) * 8 + (128 - rounding)) / 256), 0, 255); | ||
221 : | dst1[4 * dst1_stride] = CLIP((((src4 + src5) * 160 - (src3 + src6) * 48 + (src2 + src7) * 24 - (src1 + src8) * 8 + (128 - rounding)) / 256), 0, 255); | ||
222 : | dst1[5 * dst1_stride] = CLIP((((src5 + src6) * 160 - (src4 + src7) * 48 + (src3 + src8) * 24 - (src2 + src8) * 8 + (128 - rounding)) / 256), 0, 255); | ||
223 : | dst1[6 * dst1_stride] = CLIP((((src6 + src7) * 160 - (src5 + src8) * 48 + (src4 + src8) * 24 - (src3 + src7) * 8 + (128 - rounding)) / 256), 0, 255); | ||
224 : | dst1[7 * dst1_stride] = CLIP((((src7 + src8) * 160 - (src6 + src8) * 48 + (src5 + src7) * 24 - (src4 + src6) * 8 + (128 - rounding)) / 256), 0, 255); | ||
225 : | |||
226 : | dst1++; | ||
227 : | } | ||
228 : | } | ||
229 : | |||
230 : | void interpolate8x8_bilinear2(uint8_t *dst, uint8_t *src1, uint8_t *src2, int32_t dst_stride, int32_t src_stride, int32_t rounding) | ||
231 : | { | ||
232 : | int32_t i; | ||
233 : | |||
234 : | for(i = 0; i < 8; i++) | ||
235 : | { | ||
236 : | dst[0] = (src1[0] + src2[0] + (1 - rounding)) >> 1; | ||
237 : | dst[1] = (src1[1] + src2[1] + (1 - rounding)) >> 1; | ||
238 : | dst[2] = (src1[2] + src2[2] + (1 - rounding)) >> 1; | ||
239 : | dst[3] = (src1[3] + src2[3] + (1 - rounding)) >> 1; | ||
240 : | dst[4] = (src1[4] + src2[4] + (1 - rounding)) >> 1; | ||
241 : | dst[5] = (src1[5] + src2[5] + (1 - rounding)) >> 1; | ||
242 : | dst[6] = (src1[6] + src2[6] + (1 - rounding)) >> 1; | ||
243 : | dst[7] = (src1[7] + src2[7] + (1 - rounding)) >> 1; | ||
244 : | |||
245 : | dst += dst_stride; | ||
246 : | src1 += src_stride; | ||
247 : | src2 += 8; | ||
248 : | } | ||
249 : | } | ||
250 : | |||
251 : | void interpolate8x8_bilinear4(uint8_t *dst, uint8_t *src1, uint8_t *src2, uint8_t *src3, uint8_t *src4, int32_t stride, int32_t rounding) | ||
252 : | { | ||
253 : | int32_t i; | ||
254 : | |||
255 : | for(i = 0; i < 8; i++) | ||
256 : | { | ||
257 : | dst[0] = (src1[0] + src2[0] + src3[0] + src4[0] + (2 - rounding)) >> 2; | ||
258 : | dst[1] = (src1[1] + src2[1] + src3[1] + src4[1] + (2 - rounding)) >> 2; | ||
259 : | dst[2] = (src1[2] + src2[2] + src3[2] + src4[2] + (2 - rounding)) >> 2; | ||
260 : | dst[3] = (src1[3] + src2[3] + src3[3] + src4[3] + (2 - rounding)) >> 2; | ||
261 : | dst[4] = (src1[4] + src2[4] + src3[4] + src4[4] + (2 - rounding)) >> 2; | ||
262 : | dst[5] = (src1[5] + src2[5] + src3[5] + src4[5] + (2 - rounding)) >> 2; | ||
263 : | dst[6] = (src1[6] + src2[6] + src3[6] + src4[6] + (2 - rounding)) >> 2; | ||
264 : | dst[7] = (src1[7] + src2[7] + src3[7] + src4[7] + (2 - rounding)) >> 2; | ||
265 : | |||
266 : | dst += stride; | ||
267 : | src1 += stride; | ||
268 : | src2 += 8; | ||
269 : | src3 += 8; | ||
270 : | src4 += 8; | ||
271 : | } | ||
272 : | } |
No admin address has been configured | ViewVC Help |
Powered by ViewVC 1.0.4 |